In 3GPP LTE, SC-FDMA (Single-Carrier Frequency Division Multiple Access) is employed as an uplink communication method (see Non-patent Literature 1). In 3GPP LTE, a radio communication base station apparatus (hereinafter simply referred to as “base station”) allocates resources for uplink data to a radio communication terminal apparatus (hereinafter simply referred to as “terminal”) through a physical channel (for example, PDCCH (Physical Downlink Control Channel)).
Also, in 3GPP LTE, HARQ (Hybrid Automatic Repeat reQuest) is applied to downlink data from a base station to a terminal. Thus, a terminal feeds back to a base station a response signal showing an error detection result of downlink data. A terminal performs a CRC (Cyclic Redundancy Check) check for downlink data, feed backs ACK (Acknowledgment) if the result of CRC is OK (no error), and feed backs NACK (Negative Acknowledgment) if the result of CRC is NG (error present), as a response signal towards a base station. A terminal transmits this response signal (that is, an ACK/NACK signal) using, for example, an uplink control channel such as a PUCCH (Physical Uplink Control Channel), to a base station.
FIG. 1 shows a resource arrangement of a PUCCH in 3GPP LTE. A PUSCH (Physical Uplink Shared Channel) shown in FIG. 1 is a channel used for a terminal to transmit uplink data, and is used when a terminal transmits uplink data. As shown in FIG. 1, a PUCCH is placed for both ends of a system band, that is, resource blocks (RB: Resource Block, or PRB: Physical RB) on both ends of a system band. The PUCCH placed for both ends of a system band switches between slots. That is to say a frequency hopping is performed per slot.
As shown in FIG. 2, studies are underway to code-multiplex multiple response signals from multiple terminals by spreading with a ZAC (Zero Auto Correlation) sequence and a Walsh sequence (see Non-patent Literature 2). In FIG. 2, [W0, W1, W2, W3] represents a Walsh sequence of a sequence length of 4. As shown in FIG. 2, in a terminal, a response signal of ACK or NACK is subjected to the first spreading using a sequence which has a characteristic becoming a ZAC sequence (sequence length 12) in the frequency domain and in the time domain at first. Next, response signals after the first spreading are associated with each W0-W3 and are subjected to an IFFT (Inverse Fast Fourier Transform). A response signal spread in the frequency domain is transformed by this IFFT into a ZAC sequence which has a sequence length of 12 in the time domain. Also, a signal after the IFFT is further subjected to second spreading using a Walsh sequence (sequence length 4). Thus, one response signal is placed for four SC-FDMA symbols S0-S3. This applies to other terminals alike, and a response signal is spread using a ZAC sequence and a Walsh sequence. However, different terminals use ZAC sequences of different cyclic shift amounts in the time domain or use different Walsh sequences. Since the sequence length of a ZAC sequence in the time domain is 12 here, twelve ZAC sequences generated from the same ZAC sequence and having cyclic shift amounts 0-11, can be used. Since the sequence length of a Walsh sequence is 4, four mutually different Walsh sequences can be used. Thus, in an ideal communication environment, response signals from maximum 48 (12×4) terminals can be code-multiplexed.
As shown in FIG. 2, studies are underway to code-multiplex multiple the reference signals (a pilot signal) from multiple terminals (see Non-patent Literature 2). As shown in FIG. 2, to generate a reference signal of three symbols R0, R1 and R2 from a ZAC sequence (sequence length 12), first, the ZAC sequence is subjected to an IFFT in association with an orthogonal sequence of a sequence length of 3, [F0, F1, F2], such as a Fourier sequence. By this IFFT, a ZAC sequence which is a sequence length of 12 in the time domain is acquired. Then, a signal after the IFFT is spread using an orthogonal sequence [F0, F1, F2]. Thus, one reference signal (ZAC sequence) is allocated to three SC-FDMA symbols, R0, R1, and R2. This applies to other terminals alike, and one reference signal (a ZAC sequence) is allocated to three SC-FDMA symbols, R0, R1, and R2. However, different terminals use ZAC sequences of different cyclic shift amounts in the time domain or use different orthogonal sequences. Since the sequence length of a ZAC sequence in the time domain is 12 here, twelve ZAC sequences generated from the same ZAC sequence and having cyclic shift amounts 0-11, can be used. Since the sequence length of an orthogonal sequence is 3, three mutually different orthogonal sequences are used. Thus, in an ideal communication environment, reference signals from maximum 36 (12×3) terminals can be code-multiplexed.
As shown in FIG. 2, seven symbols of S0, S1, R0, R1, R2, S2, and S3 form one slot.
Here, the cross-correlation between ZAC sequences of different cyclic shift amounts which are generated from the same ZAC sequence, becomes almost zero. Thus, in an ideal communication environment, multiple response signals, which are spread and code-multiplexed using ZAC sequences (cyclic shift amount 0-11) of different cyclic shift amounts, can be separated in the time domain nearly without inter symbol interference by a correlation processing in a base station.
However, because of influences such as a misalignment of transmitting timing at a terminal and a delayed wave caused by multipath, multiple response signals from multiple terminals may not arrive at to a base station at the same time. For example, if transmission timing of a response signal spread using the ZAC sequence of cyclic shift amount 0 is later than the correct timing, the correlation peak of the ZAC sequence of cyclic shift amount 0 may appear in a detecting window for a ZAC sequence of cyclic shift amount 1. If a response signal which has been spread using the ZAC sequence of cyclic shift amount 0 has a delayed wave, an interference leakage due to the delayed wave may appear in a detecting window for a ZAC sequence of cyclic shift amount 1. Thus, in these cases, a ZAC sequence of cyclic shift amount 1 suffers interference from a ZAC sequence of cyclic shift amount 0. On the other hand, if transmission timing of a response signal spread using the ZAC sequence of cyclic shift amount 0 is earlier than the correct timing, the correlation peak of the ZAC sequence of cyclic shift amount 1 may appear in a detecting window for a ZAC sequence of cyclic shift amount 0. Thus, in this case, a ZAC sequence of cyclic shift amount 0 suffers interference from a ZAC sequence of cyclic shift amount 1. Thus, in this case, the separation characteristics of a response signal spread using the ZAC sequence of cyclic shift amount 0 and a response signal spread by a ZAC sequence of cyclic shift amount 1 are degraded. Thus, if ZAC sequences having mutually adjacent cyclic shift amounts are used, the separation characteristic of a response signal may be degraded.
Thus, previously, if multiple response signals are code-multiplexed by ZAC sequence spreading, a cyclic shift interval (a gap between cyclic shift amounts) of such a scale that does not produce inter symbol interference between ZAC sequences, is provided between ZAC sequences. For example, if the cyclic shift interval between ZAC sequences is defined as 2, between twelve sequences which have a sequence length of 12 and which have cyclic shift amounts 0-11, only six ZAC sequences of cyclic shift amounts 0, 2, 4, 6, 8, and 10, or of cyclic shift amounts 1, 3, 5, 7, 9, and 11, are used for the first spreading of a response signal. Thus, if a Walsh sequence which has a sequence length of 4 is used for second spreading of a response signal, response signals from maximum 24 (6×4) terminals can be code-multiplexed.
However, as shown in FIG. 2, the sequence length of an orthogonal sequence used to spread a reference signal is 3, so that only three mutually different orthogonal sequences can be used to spread a reference signal. Thus, if multiple response signals are separated using a reference signal shown in FIG. 2, only response signals from maximum 18 (6×3) terminals can be code-multiplexed. Thus, three out of four Walsh sequences which have a sequence length of 4 are enough, so that one Walsh sequence remains unused.
Also, as a PUCCH to be used to transmit to the above mentioned eighteen response signals, studies are underway to define eighteen PUCCHs (ACK #1-ACK #18 shown in FIG. 3) shown in FIG. 3. In FIG. 3, the horizontal axis represents the cyclic shift amount, and the vertical axis represents the sequence numbers of orthogonal code sequences (sequence numbers of a Walsh sequence or a Fourier sequence).
To reduce interference from other cells in PUCCH, a technology called cyclic shift hopping is discussed (see Non-patent Literature 3). Cyclic shift hopping refers to a technology of cyclically shifting, for example, the eighteen resources (ACK #1-ACK #18) shown in FIG. 3, in SC-FDMA symbol units (in FIG. 4, symbol 0, 1, 2, . . . , n), using the cell-specific cyclic shift hopping pattern as shown in FIG. 4, maintaining their correlation relationships on a cyclic shift axis and on an orthogonal code axis. As shown in FIG. 4, although the amount of cyclic shift which resources to which a certain response signal is allocated use changes per SC-FDMA symbol, the relative relationships of resources (cyclic shift amount and orthogonal code) under the same time and the same frequency are maintained in a cell, so that these eighteen resources are orthogonal to each other. This can randomize the combination of response signals to suffer severe interference from other cells, so that no longer do some terminals alone continue suffering severe interference from other cells. Generally, different ZAC sequences are allocated to different cells, so that the ZAC sequence differences between cells contribute to randomization of interference.
Meanwhile, in a PUCCH of 3GPP LTE the above mentioned response signal (an ACK/NACK signal) and also a CQI (Channel Quality Indicator) signal are multiplexed. Although a response signal is an one-symbol of information (information indicated by using one symbol) as mentioned above, a CQI signal is a five-symbol of information (information indicated by using five symbols). As shown in FIG. 5, a terminal spreads a CQI signal using a ZAC sequence which has a sequence length of 12, and performs an IFFT on the spread CQI signal, and transmits the CQI signal. By this means, a Walsh sequence is not applied for a CQI signal, so that, in a base station, a Walsh sequence cannot be used to separate a response signal and a CQI signal. Therefore, using a ZAC sequence, a base station despreads a response signal and CQI signal which have been spread by the ZAC sequences corresponding to different cyclic shifts, so that a base station can separate a response signal and CQI signal nearly without inter symbol interference.
Like response signals, studies are underway to apply cyclic shift hopping to CQI signals in SC-FDMA units, using cell specific cyclic shift hopping pattern to randomize inter-cell interference. As shown in FIG. 6, although the amount of cyclic shift which resources to which a certain CQI signal is allocated use changes per SC-FDMA symbol, the relative relationships of the amount of cyclic shift under the same time and the same frequency are maintained. Like response signals, different ZAC sequences are allocated to different cells of CQI signals, so that the ZAC sequence differences between cells contribute to randomization of interference.
In 3GPP LTE, the cyclic shift hopping patterns shown in FIG. 4 and FIG. 6 are associated with the cell IDs of base stations one by one.
Also, the standardization of LTE-advanced (hereinafter referred to as “LTE+”) has been started to realize much faster communication than 3GPP LTE. In LTE+, to improve average throughput and improve throughput of a terminal located near a cell edge, CoMP transmission/reception (Coordinated Multipoint Transmission/Reception) where multiple base stations cooperate to transmit and receive signals and coordinate inter-cell interference, is discussed.